U.S. patent application number 15/854105 was filed with the patent office on 2018-06-28 for calculating a four dimensional dsa dataset with variable spatial resolution.
The applicant listed for this patent is SIEMENS HEALTHCARE GMBH. Invention is credited to SONJA GEHRISCH, MARKUS KOWARSCHIK, CHRISTOPHER ROHKOHL, KEVIN ROYALTY, SEBASTIAN SCHAFER.
Application Number | 20180182132 15/854105 |
Document ID | / |
Family ID | 60327077 |
Filed Date | 2018-06-28 |
United States Patent
Application |
20180182132 |
Kind Code |
A1 |
KOWARSCHIK; MARKUS ; et
al. |
June 28, 2018 |
CALCULATING A FOUR DIMENSIONAL DSA DATASET WITH VARIABLE SPATIAL
RESOLUTION
Abstract
A method calculates a four-dimensional DSA dataset from x-ray
datasets. Each of the x-ray datasets contains a two-dimensional
x-ray projection of an examination volume in relation to a
direction of projection and a recording time. A first
three-dimensional DSA dataset of a first reconstruction volume is
determined based on the x-ray datasets. The first reconstruction
volume is a part of the examination volume. A second
three-dimensional DSA dataset of a second reconstruction volume is
determined based on the x-ray datasets. The second reconstruction
volume is a part of the first reconstruction volume. The second
three-dimensional DSA dataset is segmented. The x-ray datasets are
normalized based on the first three-dimensional DSA dataset. A
four-dimensional DSA dataset is calculated by back projection of
the normalized x-ray datasets onto the segmented second
three-dimensional DSA dataset. The four-dimensional DSA dataset
contains a number of third three-dimensional DSA datasets and
associated time information.
Inventors: |
KOWARSCHIK; MARKUS;
(NUERNBERG, DE) ; GEHRISCH; SONJA; (NUERNBERG,
DE) ; ROYALTY; KEVIN; (FITCHBURG, WI) ;
SCHAFER; SEBASTIAN; (MADISON, WI) ; ROHKOHL;
CHRISTOPHER; (BRIXEN IM THALE, AT) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
SIEMENS HEALTHCARE GMBH |
Erlangen |
|
DE |
|
|
Family ID: |
60327077 |
Appl. No.: |
15/854105 |
Filed: |
December 26, 2017 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06T 2207/10076
20130101; G06T 2207/30101 20130101; G06T 2211/404 20130101; A61B
6/504 20130101; G06T 11/006 20130101; G06T 2207/20224 20130101;
G06T 3/4007 20130101; G06T 5/50 20130101 |
International
Class: |
G06T 11/00 20060101
G06T011/00; G06T 3/40 20060101 G06T003/40; G06T 5/50 20060101
G06T005/50; A61B 6/00 20060101 A61B006/00 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 23, 2016 |
DE |
10 2016 226 195.9 |
Claims
1. A method for calculating a four-dimensional digital subtraction
angiography (DSA) dataset from x-ray datasets, which comprises the
following method steps of: receiving the x-ray datasets relating to
an examination volume by means of an interface, each of the x-ray
datasets containing a two-dimensional x-ray projection of the
examination volume in relation to a direction of projection and a
recording time of the two-dimensional x-ray projection; determining
a first three-dimensional DSA dataset of a first reconstruction
volume on a basis of the x-ray datasets by means of a calculation
unit, the first reconstruction volume is a part of the examination
volume or is identical to the examination volume; determining a
second three-dimensional DSA dataset of a second reconstruction
volume on a basis of the x-ray datasets by the calculation unit,
the second reconstruction volume is a part of the first
reconstruction volume; segmenting the second three-dimensional DSA
dataset by means of the calculation unit; normalizing the x-ray
datasets on a basis of the first three-dimensional DSA dataset by
means of the calculation unit; and calculating the four-dimensional
DSA dataset by back projection of normalized x-ray datasets onto a
segmented second three-dimensional DSA dataset by means of the
calculation unit, the four-dimensional DSA dataset having a number
of third three-dimensional DSA datasets as well as associated time
information.
2. The method according to claim 1, which further comprises
calculating each of the third three-dimensional DSA datasets of the
four-dimensional DSA dataset by back projection from precisely one
of the x-ray datasets being two-dimensional datasets, and wherein
the associated time information corresponds to the recording time
of the one two-dimensional x-ray dataset.
3. The method according to claim 1, which further comprises the
following method steps, each carried out by means of the
calculation unit: determining a confidence value for at least one
first pixel of at least one of a plurality of two dimensional x-ray
projections on a basis of the first three-dimensional DSA dataset;
assigning the confidence value to a first voxel of at least one of
the third three-dimensional DSA datasets, wherein a value of the
first voxel is based on a value of the first pixel; and
interpolating the four-dimensional DSA dataset on a basis of the
confidence value.
4. The method according to claim 3, wherein the confidence value of
the first pixel of the two-dimensional x-ray projection falls
monotonously with a plurality of vessel sections in the first
three-dimensional DSA dataset projected onto the first pixel,
wherein the vessel sections are projected in a direction of
projection of the two-dimensional x-ray projection.
5. The method according to claim 4, wherein an interpolation
relates to voxels to which a confidence value smaller than a
threshold value is assigned.
6. The method according to claim 1, wherein the first and the
second three-dimensional DSA dataset contain homogeneous voxels in
each case.
7. The method according to claim 6, wherein an orientation of the
homogeneous voxels in the first three-dimensional DSA dataset
corresponds to an orientation of the homogeneous voxels in the
second three-dimensional DSA dataset.
8. The method according to claim 6, wherein a length of edges of
the homogeneous voxels of the second three-dimensional DSA dataset
parallel in relation to a first coordinate axis is smaller than a
length of edges of the homogeneous voxels of the first
three-dimensional DSA dataset parallel in relation to the first
coordinate axis.
9. The method according to claim 6, wherein a number of the
homogeneous voxels in the first three-dimensional DSA dataset is
equal to a number of the homogeneous voxels in the second
three-dimensional DSA dataset.
10. The method according to claim 6, wherein edge lengths of the
homogeneous voxels of the second three-dimensional DSA dataset are
greater than an edge length of pixels of the two-dimensional x-ray
datasets.
11. A digital subtraction angiography (DSA) calculation unit for
calculating a four-dimensional DSA dataset, comprising: an
interface embodied for receiving x-ray datasets relating to an
examination volume, wherein each of the x-ray datasets having a
two-dimensional x-ray projection of an examination volume in
relation to a direction of projection and a recording time of the
two-dimensional x-ray projection; and a calculation apparatus
embodied for determining a first three-dimensional DSA dataset of a
first reconstruction volume on a basis of the x-ray datasets,
wherein a first reconstruction volume is a part of the examination
volume or is identical with the examination volume, said
calculation apparatus further embodied for determining a second
three-dimensional DSA dataset of a second reconstruction volume on
a basis of the x-ray datasets, wherein the second reconstruction
volume is a part of the first reconstruction volume, said
calculation apparatus further embodied for segmenting the second
three-dimensional DSA dataset, for normalizing of the x-ray
datasets on a basis of the first three-dimensional DSA dataset, and
calculating the four-dimensional DSA dataset by back projection of
normalized x-ray datasets onto a segmented second three-dimensional
DSA dataset, wherein the four-dimensional DSA dataset contains a
plurality of third three-dimensional DSA datasets as well as
associated time information.
12. The DSA calculation unit according to claim 11, wherein said
calculation apparatus calculating each of the third
three-dimensional DSA datasets of the four-dimensional DSA dataset
by back projection from precisely one of the x-ray datasets being
two-dimensional x-ray datasets, and wherein the associated time
information corresponds to a recording time of a two-dimensional
x-ray dataset.
13. An x-ray unit for recording x-ray datasets, comprising: a
digital subtraction angiography calculation unit according to claim
10.
14. A non-transitory computer-readable storage medium comprising
executable instructions to be read and executed by a digital
subtraction angiography (DSA) calculation unit which comprises the
steps of: receiving x-ray datasets relating to an examination
volume by means of an interface, each of the x-ray datasets
containing a two-dimensional x-ray projection of the examination
volume in relation to a direction of projection and a recording
time of the two-dimensional x-ray projection; determining a first
three-dimensional DSA dataset of a first reconstruction volume on a
basis of the x-ray datasets by means of a calculation apparatus,
the first reconstruction volume is a part of the examination volume
or is identical to the examination volume; determining a second
three-dimensional DSA dataset of a second reconstruction volume on
a basis of the x-ray datasets by the calculation apparatus, the
second reconstruction volume is a part of the first reconstruction
volume; segmenting the second three-dimensional DSA dataset by
means of the calculation apparatus; normalizing the x-ray datasets
on a basis of the first three-dimensional DSA dataset by means of
the calculation apparatus; and calculating a four-dimensional DSA
dataset by back projection of normalized x-ray datasets onto a
segmented second three-dimensional DSA dataset by means of the
calculation apparatus, the four-dimensional DSA dataset having a
plurality of third three-dimensional DSA datasets as well as
associated time information.
Description
CROSS-REFERENCE TO RELATED APPLICATION
[0001] This application claims the benefit, under 35 U.S.C. .sctn.
119, of German patent application DE 10 2016 226 195.9, filed Dec.
23, 2016; the prior application is herewith incorporated by
reference in its entirety.
BACKGROUND OF THE INVENTION
Field of the Invention
[0002] In digital subtraction angiography (abbreviated to DSA) one
or more vessels are shown by x-ray recordings. To suppress further
structures in the examination volume, images of a vessel on its own
are combined with images of the vessel including a contrast medium
that is to be found in the vessel. The contrast medium is
introduced into the vessel here during the examination in order to
determine parameters, in particular hydrodynamic parameters of a
fluid, wherein the fluid flows in the vessel.
[0003] In four-dimensional DSA a time-resolved series of
three-dimensional DSA image data is provided by an image
reconstruction method. Normalized two-dimensional x-ray projections
of an examination volume are back projected here together with time
information into a volume element. The two-dimensional x-ray
projections usually originate here from a rotating recording
protocol of a C-arm x-ray arc.
[0004] The multiplicative back projection is subject to
restrictions if a number of vessels or a number of vessel sections
overlap in the two-dimensional x-ray projections. In this case it
is not evident from a single x-ray projection as to the overlapping
vessel to which an x-ray signal, in particular an intensity value
or an x-ray absorption coefficient, must be assigned. This is the
case in particular when overlaps occur with vessels outside the
reconstruction volume.
[0005] The number of voxels in a three-dimensional image dataset is
defined on the basis of standards in medical imaging such as for
example digital imaging and communications in medicine (DICOM),
e.g. as 256.times.256.times.256 or as 512.times.512.times.512.
[0006] It is known that multiplicative back projection can be
carried out for a maximum size of reconstruction volume, in order
to resolve a possible overlap of vessels in the best possible way.
Because of the number of voxels defined by standards however, the
spatial resolution is thus limited by the size of the
reconstruction volume.
[0007] It is further known that multiplicative back projection can
be carried out for a smaller reconstruction volume. Although the
resolution is improved here, vessels or vessel sections outside the
reconstruction volume can falsify the results.
SUMMARY OF THE INVENTION
[0008] It is therefore the object of the present invention, taking
into account vessels outside the reconstruction volume, to achieve
a variable, in particular a finer spatial resolution of the
reconstruction volume.
[0009] The inventive manner in which the object is achieved will be
described below in relation to the claimed devices and also in
relation to the claimed method. Features, advantages or alternate
forms of embodiment mentioned here are likewise also to be
transferred to the other claimed subject matter and vice versa. In
other words the physical claims (which are directed to a device for
example) can also be further developed with features that are
described or claimed in conjunction with a method. The
corresponding functional features of the method are embodied in
such cases by corresponding physical modules.
[0010] The invention is based on x-ray datasets relating to an
examination volume being received by means of an interface, wherein
each of the x-ray datasets contains a two-dimensional x-ray
projection of the examination volume relating to a direction of
projection and a recording time of the x-ray projection. Here the
examination volume contains at least one vessel, wherein the vessel
can contain a contrast medium, and wherein the spatial density of
the contrast medium can change over time. The recording time
corresponds to the point in time of the recording of the
two-dimensional x-ray projection. A two-dimensional x-ray
projection is in particular spatially two-dimensional. The x-ray
projections can in particular also involve DSA x-ray
projections.
[0011] The invention is further based on a first three-dimensional
DSA dataset of a first reconstruction volume on the basis of the
x-ray datasets being determined by a calculation unit, wherein the
first reconstruction volume is a part of the examination volume or
is identical to the latter. A three-dimensional dataset is in
particular spatially three-dimensional.
[0012] The invention is further based on a second three-dimensional
DSA dataset of a second reconstruction volume on the basis of the
x-ray datasets being determined by the calculation unit, wherein
the second reconstruction volume is a part of the first
reconstruction volume. Here the determining of the second
three-dimensional DSA dataset is in particular only based on the
x-ray datasets, i.e. in particular not on the first
three-dimensional DSA dataset.
[0013] The invention is further based on the second
three-dimensional DSA dataset being segmented by the calculation
unit. Here the DSA dataset is segmented into at least two parts,
wherein a first part contains at least one vessel contained in the
second reconstruction volume and the inside of the vessel, and a
second part comprises the other components of the second
reconstruction volume. The first part can also comprise a number of
vessels contained in the reconstruction volume and the inside of
the vessels.
[0014] The invention is further based on the x-ray datasets being
normalized by the calculation unit on the basis of the first
three-dimensional DSA dataset.
[0015] The invention is further based on a four-dimensional DSA
dataset being calculated by back projection of the normalized x-ray
datasets onto the segmented second three-dimensional DSA dataset by
the calculation unit, wherein the four-dimensional DSA dataset
contains a number of third three-dimensional DSA datasets as well
as associated time information. Here each of the third
three-dimensional DSA datasets is assigned time information. Here
the time information belonging to a third three-dimensional DSA
dataset corresponds in particular to the time at which the state of
the vessel shown corresponds to the mapping in the
three-dimensional DSA dataset. In particular each of the voxels of
each of the third three-dimensional DSA datasets contains time
information. Time information can in particular also be a time
coordinate. A back projection can in particular be a multiplicative
back projection.
[0016] The inventors have recognized that by the first
determination of a first three-dimensional DSA dataset and by the
second determination of a second three-dimensional DSA dataset, the
normalization can be based on the first three-dimensional DSA
dataset and the segmentation and also the calculation of the
four-dimensional DSA dataset can be based on the second
three-dimensional DSA dataset. Since the first reconstruction
volume is larger than the second reconstruction volume, a variable,
in particular a better, spatial resolution of the second
three-dimensional DSA datasets and simultaneously information about
further vessels mapped in the x-ray projections from the larger
first reconstruction volume can be used.
[0017] According to a further aspect of the invention, each third
three-dimensional DSA dataset of the four-dimensional DSA dataset
is calculated by back projection from precisely one of the
two-dimensional x-ray datasets, wherein the associated time
information corresponds to the recording time of the
two-dimensional x-ray dataset. The inventors have recognized that,
through the unique assignment of an x-ray projection to a
three-dimensional DSA dataset, the time information belonging to
the three-dimensional DSA dataset can be determined uniquely and
thereby without errors.
[0018] According to a further aspect of the invention, the method
further contains a third determination of a confidence value for at
least one first pixel of at least one of the x-ray projections on
the basis of the first three-dimensional DSA dataset, an assignment
of the confidence value to a first voxel of at least one of the
third three-dimensional DSA datasets, wherein the value of the
first voxel is based on the value of the first pixel, as well as an
interpolation of the four-dimensional DSA dataset on the basis of
the confidence value. The value of the first voxel is based in
particular on the value of the first pixel, when the value of the
first voxel is calculated by back projection from the value of the
first pixel. The third determination, the assignment and the
interpolation are each carried out here by means of the calculation
unit. The inventors have recognized that it is possible by an
interpolation to establish valid and less fault-susceptible
four-dimensional DSA datasets even with vessels or vessel parts of
which the x-ray projections overlap in relation to one or more
directions of projection at one or more points in time.
[0019] According to a further possible aspect of the invention, an
interpolated intensity value of a voxel of a first of the third
three-dimensional DSA datasets is only based on the intensity
values of the corresponding voxels in the third three-dimensional
DSA datasets. In this case a corresponding voxel is in particular a
spatially corresponding voxel. The inventors have recognized that
the gradient of the concentration of a contrast medium in relation
to the time is smaller and more even than it is in relation to one
of three spatial directions, and that therefore a temporal
interpolation produces especially good results. In particular the
interpolated intensity value of the first voxel is based only on
the intensity values of one of the second voxels corresponding to
the first voxel in a second of the third three-dimensional DSA
datasets and of a third voxel corresponding to a first voxel in a
third of the third three-dimensional DSA dataset. In particular the
interpolated intensity value is determined by a linear
interpolation. The inventors have recognized that a linear
interpolation is able to be carried out especially easily and
robustly in relation to an overmatching.
[0020] According to a further aspect of the invention, the
confidence value of the first pixel of an x-ray projection falls
monotonously with a number of the vessel sections projected onto
the first pixel in the first three-dimensional DSA dataset, wherein
the vessel sections are projected in the direction of projection of
the x-ray projection. The inventors have recognized that the
reliability of back-projected three-dimensional decreases with the
number of vessel sections overlapping in an x-ray projection, and
therefore the number of the overlapping vessel sections is a
suitable criterion for the reliability or exactness of a pixel
value.
[0021] According to a further aspect of the invention, the
interpolation relates to voxels to which a confidence value smaller
than a threshold value is assigned. The inventors have recognized
that voxels for which, because of overlapping, no secure
information about the intensity value is possible, can be selected
especially quickly and easily on the basis of the threshold value.
The threshold value can be selected in particular so that all
voxels are interpolated with a confidence value smaller than the
maximum allocated confidence value. Through this the intensity
values of all voxels that cannot be determined solely by the
associated x-ray projection and therefore exhibit an uncertainty or
incorrect data, can be improved or corrected by interpolation.
[0022] According to a further aspect of the invention, the first
and the second three-dimensional DSA dataset each contain
homogeneous voxels. The voxels of a DSA dataset are in particular
homogeneous when all pairs of two voxels from the DSA dataset each
have the same spatial extent in relation to a first axis parallel
to a first voxel edge, in relation to a second axis parallel to a
second voxel edge and orthogonal to the first axis and in relation
to a third axis parallel to a third voxel edge and orthogonal to
the first axis and to the second axis. The voxels of a DSA dataset
can in particular also be isotropic, meaning that each voxel of a
DSA dataset has the same extent in relation to the first axis, the
second axis and the third axis. The value of a voxel can in
particular be an x-ray absorption coefficient or a binary value,
wherein the binary value designates whether a voxel belongs to a
particular structure. The inventors have recognized that the method
can be carried out especially quickly and efficiently with
homogeneous voxels, since the back projection does not have to be
matched to the geometry of individual voxels and can therefore be
calculated vectorized and in parallel.
[0023] According to a further aspect of the invention, the
orientation of the voxels in the first three-dimensional DSA
dataset corresponds to the orientation of the voxels in the second
three-dimensional DSA dataset. The orientation of a first voxel in
the first three-dimensional DSA dataset corresponds in particular
to the orientation of a second voxel in the second
three-dimensional DSA dataset, if each edge of the first voxel is
parallel to an edge of the second voxel. The inventors have
recognized that calculation simplifications, which allow an
especially fast and simple execution of the method, are produced in
the method by the same orientation of the voxels.
[0024] According to a further aspect of the invention, the length
of the edges of the voxels of the second three-dimensional DSA
dataset parallel in relation to a first coordinate axis is smaller
than the length of the edges of the voxels of the first
three-dimensional DSA dataset parallel in relation to the first
coordinate axis. In particular the length of the edges of the
voxels of the second three-dimensional DSA dataset parallel to a
second or third coordinate axis can be smaller than the length of
the edges of the voxel of the first three-dimensional DSA dataset
parallel to a second or third coordinate axis. The inventors have
recognized that the spatial resolution of the resulting
four-dimensional DSA dataset can be improved by such a ratio of the
edge lengths.
[0025] According to a further aspect of the invention the number of
the voxels in the first three-dimensional DSA dataset is equal to
the number of the voxels in the second three-dimensional DSA
dataset. The inventors have recognized that by having the same
number of voxels in the first and in the second three-dimensional
DSA dataset, both DSA datasets can be stored in a similar data
structure, in particular in a DICOM dataset. This simplifies memory
and data management and makes possible a standardized data
exchange.
[0026] According to a further aspect of the invention, the edge
lengths of the voxels of the second three-dimensional DSA dataset
are larger than the edge length of the pixels of the x-ray dataset.
The inventors have recognized that, by such a ratio of the edge
lengths, a pixel of one of the x-ray projections of the x-ray
datasets is at most assigned to a voxel of the second
three-dimensional DSA dataset. Through this the value of a voxel is
always determined precisely by the value of at least one pixel and
does not have to be determined by an interpolation.
[0027] The invention further relates to a DSA calculation unit for
calculating a four-dimensional DSA dataset. The DSA calculation
unit contains an interface and a calculation unit. The interface is
embodied for receiving x-ray datasets relating to an examination
volume. Each of the x-ray datasets includes a two-dimensional x-ray
projection of the examination volume in relation to a direction of
projection and a recording time of the x-ray projection. The
calculation unit is embodied for a first determination of a first
three-dimensional DSA dataset of a first reconstruction volume on
the basis of the x-ray datasets. The first reconstruction volume is
a part of the examination volume or is identical to the volume. The
calculation unit is furthermore embodied for a second determination
of a second three-dimensional DSA dataset of a second
reconstruction volume on the basis of the x-ray datasets. The
second reconstruction volume is a part of the first reconstruction
volume. The calculation unit is still furthermore embodied for the
segmentation of the second three-dimensional DSA dataset and for
normalization of the x-ray datasets on the basis of the first
three-dimensional DSA dataset. The calculation unit is embodied for
calculation of a four-dimensional DSA dataset by back projection of
the normalized x-ray datasets to the segmented second
three-dimensional DSA dataset. The four-dimensional DSA dataset
contains a number of third three-dimensional DSA datasets as well
as associated time information.
[0028] Such a DSA calculation unit can be embodied in particular to
carry out the previously described inventive method and its
aspects. The DSA calculation unit is embodied to carry out this
method and its aspects in that the interface and the calculation
unit are embodied to carry out the corresponding method steps. The
invention further relates to an x-ray unit, embodied for recording
of x-ray datasets and also comprising an inventive DSA calculation
unit.
[0029] The invention also relates to a computer program product
with a computer program as well as a computer-readable medium. A
largely software-based realization has the advantage that DSA
calculation units already used previously can be upgraded in a
simple manner in order to operate in the inventive method. Such a
computer program product, as well as the computer program, can
possibly comprise additional elements such as e.g. documentation
and/or additional components and can also have hardware components,
such as e.g. hardware keys (dongles etc.) for use of the
software.
[0030] An x-ray projection is a two-dimensional projection of an
examination volume by means of x-rays in a direction of projection,
which in particular can comprise a number of pixels. In this case
each pixel is allocated an x-ray intensity value, which is a
measure for the x-ray intensity encountered in this pixel. The
incident x-ray intensity depends on the number, the size, the shape
and the material of the objects to be found in the examination
volume. An edge length of an edge of a pixel is the length in the
examination volume, which corresponds to the edge of the pixel.
[0031] A DSA x-ray projection of an examination volume can be
determined from a first x-ray projection and a second x-ray
projection of the examination volume, wherein the first x-ray
projection and the second x-ray projection have been recorded in
relation to the same direction of projection, and wherein, at the
time of the recording of the first x-ray projection, a contrast
medium distribution other than that present at the time of the
recording of the second x-ray projection has been present in the
examination volume. The DSA x-ray projection can then be computed
from the difference between the x-ray intensities of the first
x-ray projection and the second x-ray projection.
[0032] A three-dimensional dataset of the examination volume can be
reconstructed from a number of x-ray projections from different
directions of projection. If the number of x-ray projections
involves DSA x-ray projections, a three-dimensional DSA dataset of
the examination volume can be reconstructed. A three-dimensional
dataset or a three-dimensional DSA dataset can in particular
comprise a number of voxels, to which an x-ray absorption or an
x-ray intensity is assigned. The x-ray absorption can be measured
in Hounsfield units (abbreviated to HU).
[0033] A four-dimensional DSA dataset contains a number of
three-dimensional voxels, to which time information is assigned. In
an equivalent manner a four-dimensional DSA dataset can also be
described by it containing a number of three-dimensional DSA
datasets, wherein a three-dimensional DSA dataset is assigned time
information. Time information can be acquired as time coordinates,
and the four-dimensional DSA dataset can be acquired as a time
sequence or film of three-dimensional DSA datasets.
[0034] A back projection is a method that establishes, from one or
more two-dimensional projections of a three-dimensional examination
volume, data relating to the three-dimensional examination volume.
The data relating to the three-dimensional examination volume can
in particular involve absorption coefficients or Hounsfield Units.
Since a two-dimensional projection contains less information than
the three-dimensional examination volume, further information can
be used for a back projection, for example a segmentation of the
examination volume or of a reconstruction volume.
[0035] Other features which are considered as characteristic for
the invention are set forth in the appended claims.
[0036] Although the invention is illustrated and described herein
as embodied in a calculating a four-dimensional DSA dataset with
variable spatial resolution, it is nevertheless not intended to be
limited to the details shown, since various modifications and
structural changes may be made therein without departing from the
spirit of the invention and within the scope and range of
equivalents of the claims.
[0037] The construction and method of operation of the invention,
however, together with additional objects and advantages thereof
will be best understood from the following description of specific
embodiments when read in connection with the accompanying
drawings.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING
[0038] FIG. 1 is a flow diagram of a method for calculating a
four-dimensional DSA dataset from x-ray datasets according to the
invention;
[0039] FIG. 2 is a block diagram of a DSA calculation unit;
[0040] FIG. 3 is an illustration showing an x-ray unit with the DSA
calculation unit;
[0041] FIG. 4 is a diagrammatic, perspective view of a first and a
second vessel in a first examination volume; and
[0042] FIG. 5 are x-ray datasets each containing an x-ray
projection of the first and of the second vessel.
DETAILED DESCRIPTION OF THE INVENTION
[0043] The first step of an exemplary embodiment of the method
shown is the receipt REC of x-ray datasets 500 relating to an
examination volume 400 by means of an interface 201. Each of the
x-ray datasets 500 contains a two-dimensional x-ray projection
501.1, . . . , 501.4 of the examination volume 400 in relation to a
direction of projection and a recording time of the x-ray
projection 501.1, . . . , 501.4.
[0044] In the exemplary embodiment shown the x-ray projections
501.1, . . . , 501.4 have been recorded with a C-arm x-ray device
300. Precisely two x-ray datasets 500 are recorded here for each
direction of projection, wherein no x-ray contrast medium is
present in the first vessel 403 in the first x-ray dataset from
each direction of projection in each case, and wherein x-ray
contrast medium is present in the first vessel 403 in the second
x-ray dataset in each case. In this case the x-ray datasets are
recorded without x-ray contrast medium such that the C arm 303
rotates at a predetermined angle around the examination volume 400
and x-ray projections are recorded at a constant time interval.
Furthermore the x-ray datasets 500 with x-ray contrast medium are
recorded such that the C arm 303 rotates around the examination
volume 400 at the predetermined angle and during this process x-ray
projections 501.1, . . . , 501.4 are recorded at the same constant
interval.
[0045] For each series of recordings the C arm 303 of the C-arm
x-ray device 300 rotates by 260.degree. in 12 seconds and in doing
so records 304 x-ray projections from different directions of
projection. Recording parameters containing other angles of
rotation, rotation times and numbers of projections are also
possible, in particular recording parameters such as lead to x-ray
datasets that are suitable for a three-dimensional reconstruction.
Angles of rotation that are greater than the sum of 180.degree. and
the opening angle of the x-rays of the x-ray source 301 are
particularly suitable here, in particular angles of rotation of
greater than 200.degree.. In the recording of the x-ray projections
501.1, . . . , 501.4 with contrast medium the C arm 303 can rotate
in the same circumferential direction as for the recording of the
x-ray projections without contrast medium. In this case the C arm
303 must return to its starting position between the recordings.
The C arm 303 can however also rotate in the circumferential
direction opposite to that for the recording of the x-ray
projections without contrast medium.
[0046] A two-dimensional DSA x-ray projection can then be
established from a first x-ray dataset without contrast medium and
a second x-ray dataset with contrast medium in each case, wherein
the x-ray projections of the first and of the second x-ray dataset
have been recorded from the same direction of projection, by
subtraction of the intensity values of the x-ray projection of the
second x-ray dataset from the x-ray projection of the first x-ray
dataset.
[0047] As an alternative however it is also possible in this step
of the method for x-ray datasets 400 containing DSA x-ray
projections to be received directly.
[0048] Further steps of the exemplary embodiment of the method
shown are the first determination DET-1 of a first
three-dimensional DSA dataset of a first reconstruction volume 401
on the basis of the x-ray dataset 500 by means of a calculation
apparatus 202. The first reconstruction volume 401 is a part of the
examination volume 400 or is identical to said volume. A second
determination DET-2 of a second three-dimensional DSA dataset of a
second reconstruction volume 402 on the basis of the x-ray dataset
500 by means of the calculation apparatus 202 is performed. The
second reconstruction volume 402 is part of the first
reconstruction volume 401. The second three-dimensional DSA dataset
is in particular not based on the first three-dimensional DSA
dataset. Furthermore the order in which the first determination
DET-1 and the second determination DET-2 are carried out is not
relevant.
[0049] In the exemplary embodiment shown, the first determination
DET-1 and the second determination DET-2 are each based on the
two-dimensional DSA x-ray projections established from a
subtraction. The first determination DET-1 and the second
determination DET-2 can in particular only be based on those
two-dimensional DSA-x-ray projections in which the first vessel 403
to be examined is filled completely or to a large extent with
contrast medium.
[0050] In the exemplary embodiment shown, the first
three-dimensional DSA dataset and the second three-dimensional DSA
dataset are determined by means of a cone beam reconstruction from
the two-dimensional x-ray projections or the two-dimensional DSA
x-ray projections. However other reconstruction methods are also
possible, for example a fan beam reconstruction. The Hounsfield
Units of the respective reconstruction volume are determined
here.
[0051] In the exemplary embodiment shown, the first reconstruction
volume 401 is identical to the examination volume 400. The first
reconstruction volume 401 can however also be smaller than the
examination volume 400 and be contained in the examination volume
400. The second reconstruction volume 402 is smaller than the first
reconstruction volume 401 and is part of the first reconstruction
volume 401. In the exemplary embodiment shown the first
reconstruction volume 401 and the second reconstruction volume 402
are embodied in the shape of a cube. However other geometries, in
particular a rectangular geometry, are conceivable for the
reconstruction volumes 401, 402.
[0052] In the exemplary embodiment shown, both the first
three-dimensional DSA dataset and also the second three-dimensional
DSA dataset consist of the same number of isotropic voxels, and
indeed of 512.times.512.times.512 voxels, which corresponds to the
DICOM standard. However other numbers of voxels are possible, in
particular also differing numbers of voxels are possible, in
particular also 256.times.256.times.256 voxels, which likewise
corresponds to a DICOM standard. Furthermore non-isotropic voxels
are possible. Through the same number of voxels and the different
reconstruction volume 401, 402, the second three-dimensional DSA
dataset has a better spatial resolution than the first
three-dimensional DSA dataset. Furthermore the voxels of the first
three-dimensional DSA dataset have the same alignment as the voxels
of the second three-dimensional DSA dataset. They can however also
have other alignments.
[0053] A further step of the exemplary embodiment of the method
shown is the segmentation SEG of the second DSA dataset by means of
the calculation apparatus 202. In the exemplary embodiment shown
the segmentation SEG is by means of a threshold value segmentation,
thus all voxels of the second DSA dataset are assigned with
Hounsfield Units via the threshold value of a first region, which
corresponds here to a first vessel 403, furthermore all voxels of
the second DSA dataset are assigned with Hounsfield Units below the
threshold value of a second region. However other methods are
possible for segmentation, for example region growing or active
shape models. The result of the segmentation can be expressed as
function C.sub.2, wherein the function C.sub.2 assigns a voxel with
spatial three-dimensional coordinate x to a value C.sub.2(x) if the
voxel lies in the first region, wherein the value C.sub.2(x)
corresponds to the value of the voxel in the second DSA dataset,
and wherein the function C.sub.2 assigns a voxel with the spatial
three-dimensional coordinate x to a value C.sub.2(x)=0 if the voxel
lies in the second region.
[0054] A further step of the exemplary embodiment of the method
shown is the normalization NORM of the x-ray datasets 500 on the
basis of the first three-dimensional DSA dataset by means of the
calculation apparatus 202, wherein the normalization in this
exemplary embodiment is given by the following functional
relationship:
p N ( t , u ) = p ( t , u ) .intg. L ( t , u ) I 1 ( l ) dl
##EQU00001##
[0055] Here u is a two-dimensional spatial coordinate in the
coordinate system of the x-ray detector 302, and t is a temporal
coordinate, thus in particular time information. Furthermore
I.sub.1(I) designates the intensity value of the first
three-dimensional DSA dataset at a three-dimensional spatial
coordinate I. The one-dimensional path L(t,u) corresponds to the
straight line through the point-type x-ray sources 301 and the
point u on the x-ray detector 302 at the recording time t. The path
L(t,u) is furthermore dependent on the temporal coordinate t,
because the spatial position of the x-ray source 301 and of the
x-ray detector 302 change with the temporal coordinates t. The size
p(t,u) describes the intensity value of the x-ray projection
recorded at the recording time t in the detector coordinate u. The
result p.sub.N(t,u) is the normalized intensity value of the x-ray
projection recorded at the time t in the detector coordinate u.
[0056] A further step of the exemplary embodiment of the method
shown is the calculation CALC of a four-dimensional DSA dataset by
back projection of the normalized x-ray datasets 500 to the
segmented second three-dimensional DSA dataset by the calculation
apparatus 202, wherein the four-dimensional DSA dataset contains a
number of third three-dimensional DSA datasets as well as
associated time information. In this exemplary embodiment a
multiplicative back projection is used, which is given by the
following functional relationship:
f ( t , x ) = C 2 ( x ) p N ( t , A ( t , x ) ) K * .intg. L ( t ,
A ( t , x ) ) I 2 ( l ) dl ##EQU00002##
[0057] Here x is a three-dimensional spatial coordinate and t is a
temporal coordinate, thus in particular time information. The tuple
(t,x) can therefore be expressed as a four-dimensional coordinate.
Furthermore I.sub.2(x) designates the second three-dimensional DSA
dataset. A(t,x) designates the projection of the spatial coordinate
x at recording time t to the spatial two-dimensional detector
coordinate u=A(t,x). Furthermore K designates an optional
convolution kernel, the operator * designates a convolution and
C.sub.2(x) designates the function belonging to the segmentation of
the second three-dimensional DSA dataset. Furthermore f(t,x)
designates the four-dimensional DSA dataset, which contains a
number of third three-dimensional DSA datasets x), f(t.sub.N,
x).
[0058] A further optional step of the exemplary embodiment of the
method shown is a third determination DET-3 of a confidence value
for at least one first pixel of at least one of the x-ray
projections 501.1, 501.2, 501.3, 501.4 on the basis of the first
three-dimensional DSA dataset. In the exemplary embodiment shown
the confidence value of a pixel of one of the x-ray projections
501.1, 501.2, 501.3, 501.4 is a measure for the number of the
vessels 403, 404 and/or vessel sections 403.1, 403.2, which are
mapped onto the pixel. The confidence value here is 1, if only
precisely one vessel 403, 404 and/or vessel section 403.1, 403.2 is
mapped onto the pixel, otherwise the confidence value falls with
the number of vessels 403, 404 and/or vessel sections 403.1, 403.2
mapped onto the pixel.
[0059] The confidence value of a first pixel is determined by the
one-dimensional x-ray intensity distribution along a beam starting
from the x-ray source 301 to the first pixel is established from
the first three-dimensional DSA dataset. The number of vessels 403,
404 and/or vessel sections 403.1, 403.2 projected onto the first
pixel is then the number of the vessels 403, 404 and/or vessel
sections 403.1, 403.2 through which this beam(s) passes. The number
of traversed vessels 403, 404 and/or vessel sections 403.1, 403.2
can be established on the basis of the gradients of the x-ray
intensity distribution, which shows edges of vessels 403, 404
and/or vessel sections 403.1, 403.2. The establishment of the
number of traversed vessels 403, 404 and/or vessel sections 403.1,
403.2 can further include empirical values for the average size of
a vessel 403, 404 and/or vessel section 403.1, 403.2. As an
alternative it is also possible to segment the first
three-dimensional DSA dataset and to determine the number of
traversed vessels 403, 404 and/or vessel sections 403.1, 403.2 on
the basis of the segmentation. As an alternative it is likewise
possible to determine a confidence value on the basis of the number
of voxels of the first three-dimensional DSA dataset through which
the beam passes, wherein the voxels have a value above a threshold
value.
[0060] A further optional step of the exemplary embodiment of the
method shown is an assignment ASG of the confidence value to a
first voxel of at least one of the third three-dimensional DSA
datasets, wherein the value of the first voxel is based on the
value of the first pixel, wherein the first voxel is assigned to
the first pixel via the back projection. The value of the first
voxel of one of the third three-dimensional DSA datasets is based
in this case on the value of the first pixel, if the recording time
of the x-ray projection comprising the first pixel and the time
information of the three-dimensional DSA dataset belonging to the
third three-dimensional DSA dataset belonging to the first voxel
correspond to one another, and when the beam from the x-ray source
301 to the first pixel at the recording time passes through the
first voxel. In the exemplary embodiment shown the confidence value
of each pixel of each of the x-ray projections is assigned to the
corresponding voxels of the third three-dimensional DSA datasets.
It is however also possible to undertake the assignment only for a
limited number of pixels from x-ray projections 501.1, . . . ,
501.4 or of voxels of the third three-dimensional DSA datasets.
[0061] A further optional step of the exemplary embodiment of the
method shown is an interpolation INTP of the four-dimensional DSA
dataset on the basis of the confidence value. In this case the
four-dimensional DSA dataset is interpolated in the exemplary
embodiment shown such that intensity values of those voxels of
those third three-dimensional DSA datasets to which a confidence
value smaller than a threshold value is assigned is interpolated.
In this exemplary embodiment the maximum confidence value is
selected as the threshold value, so that there is interpolation for
those voxels to which, because of overlays of vessels 403, 404
and/or vessel sections 403.1, 403.2, no exact intensity value can
be allocated. In the exemplary embodiment shown, the interpolation
for a first voxel in one of the third three-dimensional DSA
datasets is only done with reference to the intensity values of the
spatial corresponding voxels of the other of the third
three-dimensional DSA datasets, thus in particular only a temporal
interpolation. In this case two voxels correspond to one another in
particular spatially if their position is described by the same
spatial coordinates. It is however also possible to use the
intensity values of other voxels on their own or additionally, and
thereby to carry out a spatial interpolation on its own or in
addition. Furthermore, in the exemplary embodiment shown there is a
temporal linear interpolation between the intensity value of a
spatially corresponding second voxel and a spatially corresponding
third voxel as checkpoints, wherein the maximum confidence value is
assigned to the second voxel and the third voxel. Here the time
information assigned to the second voxel is smaller than the first
time information assigned to the first voxel. Furthermore there is
no other spatially corresponding voxel with a maximum confidence
value, to which time information between the second time
information and the first time information is assigned. Furthermore
the third time information assigned to the third voxel here is
larger than the first time information assigned to the first voxel.
Furthermore there is no other spatially corresponding voxel with a
maximum confidence value, to which time information between the
first time information and the third time information is assigned.
The interpolation can however be based on other or further voxels
as checkpoints, furthermore a non-linear interpolation, by means of
polynomials or polynomial trains, is also possible.
[0062] FIG. 2 shows a DSA calculation unit 200 for calculating a
four-dimensional DSA dataset. The DSA calculation unit 200 shown
here is designed to carry out an inventive method. The DSA
calculation apparatus 200 has an interface 201, a calculation
apparatus 202, a memory unit 203 and also an input and output unit
204.
[0063] The DSA calculation unit 200 can in particular involve a
computer, a microcontroller or an integrated circuit. As an
alternative the DSA calculation unit 200 can involve a real or a
virtual network of computers (a real network is referred to as a
"cluster", a virtual network is referred to as a "cloud"). The
interface 201 can involve a hardware or a software interface (for
example PCI bus, USB or Firewire). The calculation apparatus 202
can have a hardware element or software elements, for example a
microprocessor or what is known as an FPGA (Field Programmable Gate
Array). A memory unit 203 can be realized as volatile memory
(Random Access Memory, abbreviated to RAM) or as non-volatile mass
storage (hard disk, USB stick, SD card, solid state disk). The
input and output unit 204 has at least one input unit and/or at
least one output unit.
[0064] In the exemplary embodiment shown, the DSA calculation unit
200 is connected to an x-ray unit 300. The connection to the x-ray
unit 300 can however also be made by a network, for example an
intranet or the Internet. The DSA calculation unit 200 can however
also be part of the x-ray unit 300. The DSA calculation unit 200
shown here is embodied to carry out the method shown in FIG. 1, in
that the interface 201 and the calculation apparatus 202 are
embodied to carry out the respective steps of the method.
[0065] FIG. 3 shows x-ray unit 300 connected to a DSA calculation
unit 200. In the exemplary embodiment shown, the x-ray unit 300
involves a C-arm x-ray unit 300. The C-arm x-ray unit 300 has an
x-ray source 301 for emitting x-rays. The C-arm x-ray unit 300
further has an x-ray detector 302 for receiving x-rays. The x-ray
source 301 and also the x-ray detector 302 are fastened to the two
different ends of the C arm 303. The C arm 303 of the C-arm x-ray
device 300 is fastened to a pedestal 304. The pedestal 304 has
drive elements that are designed to change the position of the C
arm 303. In particular the C arm 303 can be rotated around two
different axes. The C arm x-ray device further has a control and
evaluation unit 305 and also a patient support facility 306, on
which a patient 307 can be laid. By means of the control and
evaluation unit 305 the position of the C arm 303 can be set and
the C arm 303 can be rotated around the examination volume 400.
Furthermore two-dimensional x-ray projections of the first
examination volume 400 can be recorded and evaluated by means of
the control and evaluation unit 305. As an alternative to the
exemplary embodiment shown, it is also possible for the DSA
calculation unit 200 to be embodied as a part of the control and
evaluation unit 305.
[0066] FIG. 4 shows an examination volume 400 with a first vessel
403 and a second vessel 404, wherein the examination volume further
contains a first reconstruction volume 401 and a second
reconstruction volume 402. In the exemplary embodiment shown, the
first reconstruction volume 401 is identical to the examination
volume 400, but it is also conceivable for the first reconstruction
volume 401 to be smaller than the examination volume 400 and to be
contained in the examination volume 400. Furthermore the second
reconstruction volume 402 is smaller than the first reconstruction
volume 401, and the first vessel 403 is split into a first vessel
section 403.1 and a second vessel section 403.2. The second vessel
404, in the example shown, runs outside the second reconstruction
volume 402, but within the first reconstruction volume 401.
[0067] The first reconstruction volume 401 and the second
reconstruction volume 402 are embodied here in the shape of cubes,
but other geometrical shapes of the first reconstruction volume 401
and of the second reconstruction volume 402 are also conceivable.
The edges of the cube-shaped first reconstruction volume 401 and of
the cube-shaped second reconstruction volume 402 here are parallel
to a first coordinate axis X, to a second coordinate axis Y or to a
third coordinate axis Z. The first coordinate axis X, the second
coordinate axis Y and the third coordinate axis Z, in the exemplary
embodiment shown, form a right-hand Cartesian coordinate
system.
[0068] FIG. 5 shows x-ray datasets 500 each containing one of the
four x-ray projections 501.1, 501.2, 501.3 and 501.4, which have
been recorded at different recording times. The x-ray projections
501.1, . . . , 501.4 are each projections of the examination volume
400 and thus of the first reconstruction volume 401 from different
directions. The projection of the second reconstruction volume 402
is also shown in each case in the x-ray projections 501.1, . . . ,
501.4.
[0069] The first x-ray projection 501.1 is a projection of the
examination volume 400 opposite to the direction of the first
coordinate axis X. The second x-ray projection 501.2 is a
projection of the examination volume 400 in the direction of the
second coordinate axis Y and was recorded at a time after the first
x-ray projection 501.1. The third x-ray projection 501.3 is a
projection of the examination volume 400 in the direction of the
first coordinate axis X and was recorded at a time after the second
x-ray projection 501.2. The fourth x-ray projection 501.4 is a
projection of the examination volume 400 against the direction of
the second coordinate axis Y a and was recorded at a time after the
third x-ray projection 501.3.
[0070] The x-ray projections 501.1, . . . , 501.4 each contain
projections of the first vessel 403 and of the second vessel 404.
In the second x-ray projection 501.2 and in the fourth x-ray
projection 501.4, as a result of the respective direction of
projection, the first vessel 403 and the second vessel 404 overlap
such that they cannot be displayed individually in the x-ray
projection 501.2, 501.4.
[0071] Furthermore areas 502.1, . . . , 502.4 with high intensity
are shown in the x-ray projections 501.1, . . . , 501.4. In this
exemplary embodiment the areas 502.1, . . . , 502.4 correspond to
parts of the first vessel 403 and of the second vessel 404 that are
filled with a contrast medium. Since the x-ray projections 501.1, .
. . , 501.4 were recorded at different recording times, the extent
of the areas 502.1, . . . , 502.4 with high intensity differs in
the individual x-ray projections 501.1, . . . , 501.4.
[0072] If, as is known from the prior art, only one reconstruction
volume is used for determining the four-dimensional DSA dataset,
then two options are produced in the vessel structure shown. On the
one hand the first reconstruction volume 401 can be used as the
reconstruction volume, in which the first vessel 403 as well as the
second vessel 404 is contained, and in this way the contribution of
the second vessel to the x-ray projections 501.1, . . . , 501.4 of
the x-ray datasets 500 at least by means of an interpolation can be
determined. However in this case the four-dimensional DSA dataset
has only a coarse spatial resolution. On the other hand the second
reconstruction volume 402 can be used as the reconstruction volume,
in which only the first vessel 403 is contained, and in this way
finer spatial resolution of the four-dimensional DSA dataset can be
obtained. However in this case the contribution of the second
vessel 404 to the x-ray projections 501.1, . . . , 501.4 of the
x-ray datasets 500 cannot be determined by calculation. However, in
an inventive method, it is possible to obtain both advantages
without the respective disadvantages.
* * * * *